Structure of Escherichia coli RNA polymerase holoenzyme at last

Much of the mechanistic foundations of our knowledge of regulation of gene expression at the transcriptional level have been provided by Escherichia coli and its phages. E. coli RNA polymerase (EcoRNAP) is a multisubunit enzyme composed of a catalytically active core (β′βα₂ω); subunits that are evolutionarily related to β′, β, α, and ω are present in DNA-dependent RNAPs of all organisms. In bacteria, promoter DNA sequence specificity is conferred by the dissociable factor σ, which recognizes two short DNA segments upstream of the transcription initiation start site (so called −35 and −10 elements) at promoters through interactions with two domains of its σ subunit [domains 4 (σ₄) and 2 (σ₂), respectively]. Although much information about multisubunit RNAPs has emerged from the landmark structures of the Thermus aquaticus (1) and Thermus thermophilus (2) holoenzymes, and comparisons to homologs (3–5), our detailed understanding of regulation of EcoRNAP at the molecular level has been hampered by the difficulty of producing crystals that diffract to reasonably high-resolution. Finally, 24 y after the first publication of the structure of multisubunit RNAPs and after more than 30 y of trials, this year has provided us with structures of the EcoRNAP holoenzyme. Although, and perhaps as expected, these apo structures did not reveal any major surprises, their impact is already evidenced by the elucidation of the structure of EcoRNAPσ⁷⁰ in complex with the metabolite ppGpp (6), supported by biochemical approaches (7). In PNAS, Bae et al. present the structures of EcoRNAPσ⁷⁰ in the absence and presence of an E. coli-specific inhibitor, the bacteriophage T7 gene 2 product (Gp2) (8). What could this peculiarity of bacterial and phage protein interactions possibly tell us about the universal multisubunit RNAP mechanism?

The bacteriophage T7 Gp2 is essential for phage growth by inhibiting E. coli RNAP transcription from the strong T7 early promoters, transcription that would interfere with T7 RNAP transcription of the phage late genes if it were to proceed unimpaired (9). Results of biochemical and genetic experiments indicate that Gp2 blocks EcoRNAP isomerization from its closed to its open promoter complex. Gp2 interacts with the β′ jaw domain present in the RNAP DNA channel (10), across the same surface that interacts with downstream double-stranded DNA, precluding DNA binding that would lead to open complex formation (11).

It is interesting that Gp2 does not efficiently inhibit the RNAP holoenzyme lacking the σ⁷⁰ N-terminal domain 1.1 (σ⁷⁰_∆1.1), indicating that interaction of Gp2 with the β′ jaw is not sufficient for inhibition and that must play a role in the process (11).

To elucidate the role of in the mechanism of EcoRNAPσ⁷⁰ inhibition by Gp2, Bae et al. (8) have solved four structures: RNAP SeMetσ⁷⁰, RNAP σ⁷⁰ in complex with Gp2, RNAP σ⁷⁰_∆1.1 in complex with Gp2, and RNAP σ⁷⁰_∆1.1. The initial structure (RNAP SeMetσ⁷⁰) localized to the active site channel, confirming the results of FRET and distance-constrained docking experiments (12). Although the structure contacts elements of both the β and β′ pincers, the side chain electron density within this subunit is lacking, reflecting the dynamic nature of the domain in holoenzyme (Fig. 1B). The second and third structures reveal that Gp2 contacts the β′ jaw through an extensive interface that is essentially identical, whether is absent or present [undermining a previously proposed allosteric mechanism involving Gp2 inhibition transmitted through conformational changes to the active site (11)]. The two bound structures show that Gp2 also makes contacts with the β pincer to serve as a bridge between between β and β′, which explains how Gp2 favors the closed conformation of the pincers (13). In the presence of Gp2, reorients without changes in its structural core to form a hydrophobic protein/protein interface with Gp2. Most importantly, Gp2-σ_1.1 contacts give rise to newly described σ_1.1 electron density. These contacts were confirmed through the analysis of the products of UV cross-linking between the RNAP holoenzyme with and without domain 1.1 and Gp2. Effectively, Gp2 glues inside the active site channel (Fig. 1A), misappropriating σ⁷⁰ through this interaction to block the subsequent entry of DNA into the channel to allow open complex formation (Fig. 1C).

Fig. 1.

Schematic of EcoRNAPσ⁷⁰. (A) σ_1.1 is present in the active site channel but displaced during open complex formation allowing downstream DNA binding (B). (C) EcoRNAPσ⁷⁰ is complexed with Gp2, which interacts with the β′ jaw and σ_1.1, blocking downstream DNA binding and open complex formation. (Adapted from ref. 8.) (D) Schematic representation of the yeast RNAPII 32-protein preinitiation complex. RNAPII is depicted in gray at the bottom. The general transcription factors are depicted at the top with the C terminus of Ssl2, a component of TFIIH, extending over the cleft to interact with polII and preventing the DNA from entering the cleft. (Adapted from ref. 15.)

Phages seem to regard σ as a good target for host exploitation. Bacteriophage T4 misappropriates σ⁷⁰ by binding of its AsiA protein to σ₄, effectively inhibiting the utilization of most host promoters early in infection. Through its T4 MotA protein, which interacts with AsiA bound to σ₄, T4 remodels σ specificity and redirects host RNAP to T4 middle promoters (14).

The elucidation of EcoRNAP holoenzyme structures has wide-ranging implications. These structures provide unambiguous structural information about the positioning of in the holoenzyme, and how it allows for increased specificity in the process of promoter recognition by interfering with the interaction of RNAP with promoterless DNA. Cognate promoters lead to exit of from the active site channel, allowing for open complex formation and transcription initiation, whereas nonspecific DNAs fail in this respect.

Regulation of promoter binding by multisubunit RNAPs is not restricted to bacterial transcription. Three recent reports indicate that eukaryotic RNAPs I and II might use similar mechanisms to restrict entry of the DNA to the active site channel and regulate transcription initiation. The architecture of the 32-protein preinitiation complex of yeast RNAP II with a full complement of general transcription factors (TFIIA, TFIIB, TBP, TFIIE, TFIIF, and TFIIH) has been determined by cryoEM and combinations of

Bae et al. present the structures of EcoRNAPσ⁷⁰ in the absence and presence of an E. coli-specific inhibitor, the bacteriophage T7 gene 2 product (Gp2).

chemical cross-linking and mass spectrometry (15). The complex of transcription factors sits on top of the cleft of yeast RNAPII. The DNA template is present on the surface of the complex, interacting only with the general transcription factors, reminiscent of the interaction of the −35 and −0 bacterial promoter elements with σ domains 4.2 and 2.4. DNA entry to the RNAP II cleft is blocked by the C-terminal region of Ssl2, a component of TFIIH (Fig. 1D). Fernández-Tornero et al. (16) and Engel et al. (17) have just described the crystal structure of the 14-subunit yeast RNAP I, responsible for ribosomal RNA transcription. The structures revealed two elements: the “connector,” which acts in trans by inserting itself into the cleft of another monomer, leading to the formation and stabilization of an inactive dimer; and the “expander,” present at the basis of the active cleft and overlapping with the site of binding of the template strand. Both elements are candidates for regulation of transcription initiation (16, 17). Moreover, this mode of regulation is not restricted to multisubunit DNA-dependent RNAPs. The bacteriophage N4 virion RNAP, a distantly related member of the T7-like RNAP family, is injected into the host in an inactive conformation, in which the path of template DNA into the active channel is blocked by two elements: the “plug” and the “motif B loop” (18). Binding of the DNA-hairpin promoter leads to movement of the plug and restructuring of the motif B loop into a helix to open the DNA channel, leading to activation of the polymerase for transcription initiation.

We look forward to the structural elucidation of a large number of distinctive regulatory mechanisms involving proteins that have been extensively characterized biochemically and genetically, and that interact specifically with EcoRNAP.